Stock price prediction using Long-Short Term Memory¶

Problem Statement¶

The main aim of the project is to predict close price of next 20 days

Import library and dataset¶

In [1]:
import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
In [2]:
df=pd.read_csv(r"C:\Users\dell\Downloads\ICICIBANK.NS(2).csv")
In [3]:
df.head()
Out[3]:
Date Open High Low Close Adj Close Volume
0 2007-10-29 220.018188 231.272720 217.363632 225.572723 184.000305 27834911.0
1 2007-10-30 227.500000 235.263641 218.663635 225.490906 183.933609 27964673.0
2 2007-10-31 225.636368 231.090912 220.654541 228.009094 185.987656 11675427.0
3 2007-11-01 230.909088 239.636368 225.454544 236.054550 192.550400 27922845.0
4 2007-11-02 225.927277 245.272720 224.727264 242.436356 197.756027 10804777.0
In [4]:
df.tail()
Out[4]:
Date Open High Low Close Adj Close Volume
3945 2023-10-19 936.000000 945.000000 934.000000 935.750000 935.750000 8612676.0
3946 2023-10-20 932.000000 940.000000 930.000000 932.750000 932.750000 6586576.0
3947 2023-10-23 942.000000 944.799988 925.400024 929.950012 929.950012 11856890.0
3948 2023-10-25 926.049988 933.799988 911.400024 916.099976 916.099976 12593434.0
3949 2023-10-26 914.299988 916.599976 899.000000 908.650024 908.650024 21693345.0

Check data to perform¶

In [5]:
df=df.dropna()
In [6]:
df.shape
Out[6]:
(3943, 7)
In [7]:
df.isnull().sum()
Out[7]:
Date         0
Open         0
High         0
Low          0
Close        0
Adj Close    0
Volume       0
dtype: int64
In [8]:
df.info()

<class 'pandas.core.frame.DataFrame'>
Int64Index: 3943 entries, 0 to 3949
Data columns (total 7 columns):
 #   Column     Non-Null Count  Dtype  
---  ------     --------------  -----  
 0   Date       3943 non-null   object 
 1   Open       3943 non-null   float64
 2   High       3943 non-null   float64
 3   Low        3943 non-null   float64
 4   Close      3943 non-null   float64
 5   Adj Close  3943 non-null   float64
 6   Volume     3943 non-null   float64
dtypes: float64(6), object(1)
memory usage: 246.4+ KB
In [9]:
df.describe()
Out[9]:
Open High Low Close Adj Close Volume
count 3943.000000 3943.000000 3943.000000 3943.000000 3943.000000 3.943000e+03
mean 338.973436 343.174931 334.393065 338.789758 323.359816 2.337523e+07
std 233.160139 234.756053 231.478391 233.198516 237.868090 1.942975e+07
min 47.945454 49.527271 45.872726 47.809093 39.700455 0.000000e+00
25% 182.399994 185.422722 179.663635 182.150002 158.941170 1.263095e+07
50% 254.818176 258.090912 251.000000 254.136368 240.323715 1.793725e+07
75% 391.225006 396.625000 387.024994 392.000000 384.612030 2.686832e+07
max 1004.950012 1008.700012 990.450012 998.299988 992.450012 2.868577e+08
In [10]:
sd=df.iloc[0][0]
ed=df.iloc[-1][0]

print('Starting Date',sd)
print('Ending Date',ed)

Starting Date 2007-10-29
Ending Date 2023-10-26
In [11]:
x=df.Open
y=df.Close
a=df.High
b=df.Low
In [12]:
import plotly.express as px
from itertools import cycle


names= cycle(['Stock Open Price','Stock Close Price','Stock High Price','Stock Low Price'])


fig = px.line(df, x='Date', y=["Open",'Close',"High","Low"], labels={'Date': 'Date','value':'Stock value'},
              title='IDEA.NS stock prices chart')

fig.update_layout( font_color='black',legend_title_text='Stock Parameters')
fig.for_each_trace(lambda t:  t.update(name = next(names)))



fig.show()

C:\Users\Armaan\anaconda3\lib\site-packages\scipy\__init__.py:155: UserWarning: A NumPy version >=1.18.5 and <1.25.0 is required for this version of SciPy (detected version 1.26.1
  warnings.warn(f"A NumPy version >={np_minversion} and <{np_maxversion}"

Original closing price¶

In [13]:
import plotly.express as px
from itertools import cycle
name=cycle(["Close price"])
fig=px.line(df,x="Date",y=["Close"])
fig.update_layout( font_color='black',legend_title_text='Stock Parameters')
fig.for_each_trace(lambda t:  t.update(name = next(name)))
fig.show()

Data visualisation of the data¶

In [14]:
import seaborn as sns
from scipy import stats
import scipy as scipy
import matplotlib.pyplot as plt
In [15]:
fig, axs = plt.subplots(2, 2, figsize=(18, 11))
sns.kdeplot(  df['Open'],shade=True,label='Open',color='gold',ax=axs[0, 0],alpha=0.5)
sns.kdeplot(df['Close'],shade=True,label='Close',color='green',ax=axs[0, 1],alpha=0.5)
sns.kdeplot(df['High'],shade=True,label='High',color='crimson',ax=axs[1, 0],alpha=0.5)
sns.kdeplot(df['Low'],shade=True,label='Low',color='slateblue',ax=axs[1, 1],alpha=0.5)
plt.show()

Check smoothness of data¶

In [16]:
plt.figure(figsize=(18,8))

plt.plot(x, label="Opening price")
ma50=x.rolling(50).mean()
ma100=x.rolling(100).mean()
ma150=x.rolling(150).mean()
plt.plot(ma50, "r", label="MA 50")
plt.plot(ma100, "g", label="MA 100")
plt.plot(ma150, "violet", label="MA 150")
plt.legend(loc="best")
plt.show()

plt.figure(figsize=(18,8))
plt.plot(y, label="Closing price")
ma50=y.rolling(50).mean()
ma100=y.rolling(100).mean()
ma150=y.rolling(150).mean()
plt.plot(ma50, "r",label="MA 50")
plt.plot(ma100, "g", label="MA 100")
plt.plot(ma150, "violet",label="MA 150")
plt.legend(loc="best")
plt.show()
In [17]:
plt.figure(figsize=(18,10))

plt.plot(a, label="High price")
ma50=a.rolling(50).mean()
ma100=a.rolling(100).mean()
ma150=a.rolling(150).mean()
plt.plot(ma50, "r", label="MA 50")
plt.plot(ma100, "g", label="MA 100")
plt.plot(ma150, "violet", label="MA 150")
plt.legend(loc="best")
plt.show()

plt.figure(figsize=(18,10))
plt.plot(b, label="Low price")
ma50=b.rolling(50).mean()
ma100=b.rolling(100).mean()
ma150=b.rolling(150).mean()
plt.plot(ma50, "r", label="MA 50")
plt.plot(ma100, "g", label="MA 100")
plt.plot(ma150, "violet", label="MA 150")
plt.legend(loc="best")
plt.show()

Insights¶

  • A moving average is a statistic that captures the average change in a data series over time.
  • I plotted moving averages for 50, 150, and 100 periods alongside the original line plot. However, I chose to use the 50-period moving average for further analysis because it provides a smoother representation.
In [18]:
df1=df.reset_index()['Close']
In [19]:
df1
Out[19]:
0       225.572723
1       225.490906
2       228.009094
3       236.054550
4       242.436356
           ...    
3938    935.750000
3939    932.750000
3940    929.950012
3941    916.099976
3942    908.650024
Name: Close, Length: 3943, dtype: float64

Scaling using MinMaxScaler¶

In [20]:
from sklearn.preprocessing import MinMaxScaler
Scaler=MinMaxScaler(feature_range=(0,1))
df1=Scaler.fit_transform(np.array(df1).reshape(-1,1))  #Fit to data, then transform it.
In [21]:
df1
Out[21]:
array([[0.18702297],
       [0.18693689],
       [0.18958625],
       ...,
       [0.92808982],
       [0.91351836],
       [0.90568035]])
In [22]:
df1.shape
Out[22]:
(3943, 1)

Split the data into train and test¶

In [23]:
training_size=int(len(df1)*0.75)
test_size=len(df1)-training_size
train_data, test_data=df1[0:training_size,:], df1[training_size:len(df1),:1]
#time_step -> the len of the input sequence
In [24]:
training_size, test_size
Out[24]:
(2957, 986)
In [25]:
import numpy
def create_dataset(dataset, time_step=1): 
    dataX, dataY=[], []
    for i in range(len(dataset)-time_step-1):
        a=dataset[i:(i+time_step), 0]
        dataX.append(a)
        dataY.append(dataset[i+time_step, 0])
    return numpy.array(dataX), numpy.array(dataY)
In [26]:
time_step=50
X_train, y_train=create_dataset(train_data, time_step)
X_test, ytest=create_dataset(test_data, time_step)
In [27]:
print(X_train.shape), print(y_train.shape)

(2906, 50)
(2906,)
Out[27]:
(None, None)
In [28]:
print(X_test.shape), print(ytest.shape)

(935, 50)
(935,)
Out[28]:
(None, None)
In [29]:
X_train=X_train.reshape(X_train.shape[0], X_train.shape[1], 1)
X_test=X_test.reshape(X_test.shape[0], X_test.shape[1], 1)
In [30]:
print("X_train: ", X_train.shape)
print("X_test: ", X_test.shape)

X_train:  (2906, 50, 1)
X_test:  (935, 50, 1)

Model training¶

In [47]:
from tensorflow.keras.models import Sequential
from tensorflow.keras.layers import Dense
from tensorflow.keras.layers import LSTM
import tensorflow as tf
In [48]:
tf.__version__
Out[48]:
'2.14.0'
In [49]:
model=Sequential()
model.add(LSTM(50, return_sequences=True, input_shape=(50, 1)))
model.add(LSTM(50, return_sequences=True))
model.add(LSTM(50))
model.add(Dense(1))
model.compile(loss='mean_squared_error', optimizer='adam')

Insights¶

  • Here we start to build the model.
  • Now, we add four layers to the model. First can remember patterns in a sequences of data and set ups to look at time_step number of data points at a time.
  • Second layer takes the output of the previous layer as its input and Third layer only returns the final output not the whole sequence.
  • Fourth layer connect everything together. It takes output from the previous layer and transform it into the final prediction
  • Here adam is used to improve the model and reduce the loss.
In [50]:
model.summary()

Model: "sequential_1"
_________________________________________________________________
 Layer (type)                Output Shape              Param #   
=================================================================
 lstm_3 (LSTM)               (None, 50, 50)            10400     
                                                                 
 lstm_4 (LSTM)               (None, 50, 50)            20200     
                                                                 
 lstm_5 (LSTM)               (None, 50)                20200     
                                                                 
 dense_1 (Dense)             (None, 1)                 51        
                                                                 
=================================================================
Total params: 50851 (198.64 KB)
Trainable params: 50851 (198.64 KB)
Non-trainable params: 0 (0.00 Byte)
_________________________________________________________________
In [51]:
model.fit(X_train, y_train, validation_data=(X_test, ytest), epochs=50, batch_size=64, verbose=1)

Epoch 1/50
46/46 [==============================] - 15s 157ms/step - loss: 0.0043 - val_loss: 0.0046
Epoch 2/50
46/46 [==============================] - 6s 130ms/step - loss: 2.2950e-04 - val_loss: 0.0018
Epoch 3/50
46/46 [==============================] - 6s 135ms/step - loss: 2.0336e-04 - val_loss: 0.0013
Epoch 4/50
46/46 [==============================] - 6s 124ms/step - loss: 1.9159e-04 - val_loss: 0.0012
Epoch 5/50
46/46 [==============================] - 6s 136ms/step - loss: 1.9402e-04 - val_loss: 0.0013
Epoch 6/50
46/46 [==============================] - 6s 135ms/step - loss: 1.8168e-04 - val_loss: 0.0011
Epoch 7/50
46/46 [==============================] - 6s 132ms/step - loss: 1.6864e-04 - val_loss: 0.0010
Epoch 8/50
46/46 [==============================] - 6s 123ms/step - loss: 1.6761e-04 - val_loss: 0.0010
Epoch 9/50
46/46 [==============================] - 6s 132ms/step - loss: 1.6540e-04 - val_loss: 9.0087e-04
Epoch 10/50
46/46 [==============================] - 6s 137ms/step - loss: 1.5382e-04 - val_loss: 8.6727e-04
Epoch 11/50
46/46 [==============================] - 6s 128ms/step - loss: 1.4866e-04 - val_loss: 8.0444e-04
Epoch 12/50
46/46 [==============================] - 7s 151ms/step - loss: 1.5487e-04 - val_loss: 8.2320e-04
Epoch 13/50
46/46 [==============================] - 7s 148ms/step - loss: 1.6304e-04 - val_loss: 0.0017
Epoch 14/50
46/46 [==============================] - 7s 146ms/step - loss: 1.3250e-04 - val_loss: 7.1270e-04
Epoch 15/50
46/46 [==============================] - 7s 152ms/step - loss: 1.3169e-04 - val_loss: 7.0229e-04
Epoch 16/50
46/46 [==============================] - 7s 149ms/step - loss: 1.3150e-04 - val_loss: 0.0011
Epoch 17/50
46/46 [==============================] - 7s 149ms/step - loss: 1.3041e-04 - val_loss: 0.0010
Epoch 18/50
46/46 [==============================] - 7s 151ms/step - loss: 1.2801e-04 - val_loss: 6.7950e-04
Epoch 19/50
46/46 [==============================] - 7s 148ms/step - loss: 1.2404e-04 - val_loss: 0.0011
Epoch 20/50
46/46 [==============================] - 7s 148ms/step - loss: 1.2942e-04 - val_loss: 0.0010
Epoch 21/50
46/46 [==============================] - 7s 146ms/step - loss: 1.1498e-04 - val_loss: 0.0012
Epoch 22/50
46/46 [==============================] - 7s 143ms/step - loss: 1.0784e-04 - val_loss: 7.6118e-04
Epoch 23/50
46/46 [==============================] - 7s 145ms/step - loss: 1.0746e-04 - val_loss: 4.9164e-04
Epoch 24/50
46/46 [==============================] - 7s 148ms/step - loss: 1.1047e-04 - val_loss: 5.8763e-04
Epoch 25/50
46/46 [==============================] - 7s 148ms/step - loss: 9.9274e-05 - val_loss: 5.0748e-04
Epoch 26/50
46/46 [==============================] - 7s 149ms/step - loss: 9.7903e-05 - val_loss: 4.3017e-04
Epoch 27/50
46/46 [==============================] - 7s 149ms/step - loss: 9.3957e-05 - val_loss: 4.1561e-04
Epoch 28/50
46/46 [==============================] - 7s 149ms/step - loss: 9.0603e-05 - val_loss: 4.0638e-04
Epoch 29/50
46/46 [==============================] - 7s 151ms/step - loss: 8.8263e-05 - val_loss: 7.5524e-04
Epoch 30/50
46/46 [==============================] - 7s 145ms/step - loss: 8.9302e-05 - val_loss: 7.7277e-04
Epoch 31/50
46/46 [==============================] - 7s 143ms/step - loss: 9.5109e-05 - val_loss: 5.7587e-04
Epoch 32/50
46/46 [==============================] - 7s 148ms/step - loss: 8.4691e-05 - val_loss: 3.9402e-04
Epoch 33/50
46/46 [==============================] - 7s 145ms/step - loss: 7.9223e-05 - val_loss: 6.4847e-04
Epoch 34/50
46/46 [==============================] - 7s 148ms/step - loss: 9.0784e-05 - val_loss: 3.4336e-04
Epoch 35/50
46/46 [==============================] - 7s 147ms/step - loss: 8.0083e-05 - val_loss: 4.3051e-04
Epoch 36/50
46/46 [==============================] - 6s 135ms/step - loss: 7.9593e-05 - val_loss: 6.1323e-04
Epoch 37/50
46/46 [==============================] - 7s 147ms/step - loss: 7.9023e-05 - val_loss: 3.4364e-04
Epoch 38/50
46/46 [==============================] - 7s 148ms/step - loss: 6.7289e-05 - val_loss: 4.0019e-04
Epoch 39/50
46/46 [==============================] - 7s 145ms/step - loss: 6.5826e-05 - val_loss: 4.5065e-04
Epoch 40/50
46/46 [==============================] - 7s 144ms/step - loss: 7.5878e-05 - val_loss: 2.5908e-04
Epoch 41/50
46/46 [==============================] - 7s 144ms/step - loss: 6.3315e-05 - val_loss: 2.5405e-04
Epoch 42/50
46/46 [==============================] - 7s 145ms/step - loss: 6.5623e-05 - val_loss: 2.6985e-04
Epoch 43/50
46/46 [==============================] - 7s 146ms/step - loss: 6.0867e-05 - val_loss: 2.3948e-04
Epoch 44/50
46/46 [==============================] - 7s 146ms/step - loss: 6.3305e-05 - val_loss: 3.6197e-04
Epoch 45/50
46/46 [==============================] - 7s 143ms/step - loss: 5.9484e-05 - val_loss: 6.1750e-04
Epoch 46/50
46/46 [==============================] - 6s 141ms/step - loss: 6.1605e-05 - val_loss: 5.3662e-04
Epoch 47/50
46/46 [==============================] - 7s 142ms/step - loss: 5.7136e-05 - val_loss: 4.6437e-04
Epoch 48/50
46/46 [==============================] - 7s 145ms/step - loss: 5.1833e-05 - val_loss: 2.6112e-04
Epoch 49/50
46/46 [==============================] - 7s 144ms/step - loss: 5.1574e-05 - val_loss: 5.4248e-04
Epoch 50/50
46/46 [==============================] - 7s 146ms/step - loss: 5.9817e-05 - val_loss: 2.2339e-04
Out[51]:
<keras.src.callbacks.History at 0x2bf02ed9490>
In [61]:
train_predict=model.predict(X_train)
test_predict=model.predict(X_test)

91/91 [==============================] - 3s 30ms/step
30/30 [==============================] - 1s 32ms/step
In [62]:
train_predict=Scaler.inverse_transform(train_predict)
test_predict=Scaler.inverse_transform(test_predict)
In [63]:
import math
In [64]:
from sklearn.metrics import mean_squared_error
math.sqrt(mean_squared_error(y_train, train_predict))
Out[64]:
241.77297467567846
In [65]:
import math
math.sqrt(mean_squared_error(ytest, test_predict))
Out[65]:
723.0922327730151
In [66]:
plt.figure(figsize=(18,10))
look_back=50
trainPredictPlot=numpy.empty_like(df1)
trainPredictPlot[:, :]=np.nan
trainPredictPlot[look_back:len(train_predict)+look_back, :]=train_predict

testPredictPlot=numpy.empty_like(df1)
testPredictPlot[:, :]=np.nan
testPredictPlot[len(train_predict)+(look_back*2)+1:len(df1)-1, :]=test_predict

plt.plot(Scaler.inverse_transform(df1))
plt.plot(trainPredictPlot)
plt.plot(testPredictPlot)
plt.show()
In [67]:
len(test_data)
Out[67]:
986

Insights¶

  • Blue color shows orinal closing price.
  • Here Orange color shows predicted training data and green color shows predicted test data.
In [72]:
x_input=test_data[936:].reshape(1,-1)
x_input.shape
Out[72]:
(1, 50)
In [73]:
temp_input=list(x_input)
temp_input=temp_input[0].tolist()
In [74]:
temp_input
Out[74]:
[0.9592315926392961,
 0.9560227370720894,
 0.9509201358525375,
 0.9498680479206484,
 0.9546550353856887,
 0.9512883413785883,
 0.9669118219170315,
 0.9691212444491643,
 0.9706467845754586,
 0.9708045598900764,
 0.9678587263058421,
 0.9585477733587336,
 0.95838993491884,
 0.968858222466192,
 0.967437865882976,
 0.9685426087116805,
 0.9570748250039786,
 0.9656493668989853,
 0.9708045598900764,
 0.9789056495906779,
 0.9917937267563198,
 0.9910572525789422,
 0.9888478931720855,
 0.9938453108485589,
 0.9914781130018083,
 0.9882692574346016,
 0.9589685706563237,
 0.9494997792693216,
 0.9558648986321957,
 0.9478164638284093,
 0.9419247335346647,
 0.9413986895687201,
 0.9511831578355098,
 0.9389788999504303,
 0.9281949986485667,
 0.9395575356879141,
 0.9441341560667974,
 0.9371903378411636,
 0.9504466836581321,
 0.952445612853556,
 0.9536555392253389,
 0.9505518672012107,
 0.9506571138695653,
 0.9532873336992881,
 0.9426086159405029,
 0.9341919124853899,
 0.9310356486897226,
 0.9280898151054882,
 0.9135183593736581,
 0.9056803547813048]
In [75]:
from numpy import array

list_output=[]
n_step=50
i=0
while(i<20):
    
    if(len(temp_input)>50):
        
        x_input=np.array(temp_input[1:])
        print("{} day input {}".format(i,x_input))
        x_input=x_input.reshape(1,-1)
        x_input=x_input.reshape((1,n_step,1))
        
        yhat=model.predict(x_input,verbose=0)
        print("{} day input {}".format(i,yhat))
        temp_input.extend(yhat[0].tolist())
        temp_input=temp_input[1:]
        
        list_output.extend(yhat.tolist())
        i=i+1
    else:
        x_input=x_input.reshape((1,n_step,1))
        yhat=model.predict(x_input,verbose=0)
        print(yhat[0])
        
        temp_input.extend(yhat[0].tolist())
        print(len(temp_input))
        list_output.extend(yhat.tolist())
        i=i+1
        
        
print(list_output)
    

[0.9103341]
51
1 day input [0.95602274 0.95092014 0.94986805 0.95465504 0.95128834 0.96691182
 0.96912124 0.97064678 0.97080456 0.96785873 0.95854777 0.95838993
 0.96885822 0.96743787 0.96854261 0.95707483 0.96564937 0.97080456
 0.97890565 0.99179373 0.99105725 0.98884789 0.99384531 0.99147811
 0.98826926 0.95896857 0.94949978 0.9558649  0.94781646 0.94192473
 0.94139869 0.95118316 0.9389789  0.928195   0.93955754 0.94413416
 0.93719034 0.95044668 0.95244561 0.95365554 0.95055187 0.95065711
 0.95328733 0.94260862 0.93419191 0.93103565 0.92808982 0.91351836
 0.90568035 0.91033411]
1 day input [[0.9136142]]
2 day input [0.95092014 0.94986805 0.95465504 0.95128834 0.96691182 0.96912124
 0.97064678 0.97080456 0.96785873 0.95854777 0.95838993 0.96885822
 0.96743787 0.96854261 0.95707483 0.96564937 0.97080456 0.97890565
 0.99179373 0.99105725 0.98884789 0.99384531 0.99147811 0.98826926
 0.95896857 0.94949978 0.9558649  0.94781646 0.94192473 0.94139869
 0.95118316 0.9389789  0.928195   0.93955754 0.94413416 0.93719034
 0.95044668 0.95244561 0.95365554 0.95055187 0.95065711 0.95328733
 0.94260862 0.93419191 0.93103565 0.92808982 0.91351836 0.90568035
 0.91033411 0.91361421]
2 day input [[0.9200785]]
3 day input [0.94986805 0.95465504 0.95128834 0.96691182 0.96912124 0.97064678
 0.97080456 0.96785873 0.95854777 0.95838993 0.96885822 0.96743787
 0.96854261 0.95707483 0.96564937 0.97080456 0.97890565 0.99179373
 0.99105725 0.98884789 0.99384531 0.99147811 0.98826926 0.95896857
 0.94949978 0.9558649  0.94781646 0.94192473 0.94139869 0.95118316
 0.9389789  0.928195   0.93955754 0.94413416 0.93719034 0.95044668
 0.95244561 0.95365554 0.95055187 0.95065711 0.95328733 0.94260862
 0.93419191 0.93103565 0.92808982 0.91351836 0.90568035 0.91033411
 0.91361421 0.92007852]
3 day input [[0.9262401]]
4 day input [0.95465504 0.95128834 0.96691182 0.96912124 0.97064678 0.97080456
 0.96785873 0.95854777 0.95838993 0.96885822 0.96743787 0.96854261
 0.95707483 0.96564937 0.97080456 0.97890565 0.99179373 0.99105725
 0.98884789 0.99384531 0.99147811 0.98826926 0.95896857 0.94949978
 0.9558649  0.94781646 0.94192473 0.94139869 0.95118316 0.9389789
 0.928195   0.93955754 0.94413416 0.93719034 0.95044668 0.95244561
 0.95365554 0.95055187 0.95065711 0.95328733 0.94260862 0.93419191
 0.93103565 0.92808982 0.91351836 0.90568035 0.91033411 0.91361421
 0.92007852 0.92624009]
4 day input [[0.93162614]]
5 day input [0.95128834 0.96691182 0.96912124 0.97064678 0.97080456 0.96785873
 0.95854777 0.95838993 0.96885822 0.96743787 0.96854261 0.95707483
 0.96564937 0.97080456 0.97890565 0.99179373 0.99105725 0.98884789
 0.99384531 0.99147811 0.98826926 0.95896857 0.94949978 0.9558649
 0.94781646 0.94192473 0.94139869 0.95118316 0.9389789  0.928195
 0.93955754 0.94413416 0.93719034 0.95044668 0.95244561 0.95365554
 0.95055187 0.95065711 0.95328733 0.94260862 0.93419191 0.93103565
 0.92808982 0.91351836 0.90568035 0.91033411 0.91361421 0.92007852
 0.92624009 0.93162614]
5 day input [[0.9357327]]
6 day input [0.96691182 0.96912124 0.97064678 0.97080456 0.96785873 0.95854777
 0.95838993 0.96885822 0.96743787 0.96854261 0.95707483 0.96564937
 0.97080456 0.97890565 0.99179373 0.99105725 0.98884789 0.99384531
 0.99147811 0.98826926 0.95896857 0.94949978 0.9558649  0.94781646
 0.94192473 0.94139869 0.95118316 0.9389789  0.928195   0.93955754
 0.94413416 0.93719034 0.95044668 0.95244561 0.95365554 0.95055187
 0.95065711 0.95328733 0.94260862 0.93419191 0.93103565 0.92808982
 0.91351836 0.90568035 0.91033411 0.91361421 0.92007852 0.92624009
 0.93162614 0.93573272]
6 day input [[0.9381914]]
7 day input [0.96912124 0.97064678 0.97080456 0.96785873 0.95854777 0.95838993
 0.96885822 0.96743787 0.96854261 0.95707483 0.96564937 0.97080456
 0.97890565 0.99179373 0.99105725 0.98884789 0.99384531 0.99147811
 0.98826926 0.95896857 0.94949978 0.9558649  0.94781646 0.94192473
 0.94139869 0.95118316 0.9389789  0.928195   0.93955754 0.94413416
 0.93719034 0.95044668 0.95244561 0.95365554 0.95055187 0.95065711
 0.95328733 0.94260862 0.93419191 0.93103565 0.92808982 0.91351836
 0.90568035 0.91033411 0.91361421 0.92007852 0.92624009 0.93162614
 0.93573272 0.93819141]
7 day input [[0.93902886]]
8 day input [0.97064678 0.97080456 0.96785873 0.95854777 0.95838993 0.96885822
 0.96743787 0.96854261 0.95707483 0.96564937 0.97080456 0.97890565
 0.99179373 0.99105725 0.98884789 0.99384531 0.99147811 0.98826926
 0.95896857 0.94949978 0.9558649  0.94781646 0.94192473 0.94139869
 0.95118316 0.9389789  0.928195   0.93955754 0.94413416 0.93719034
 0.95044668 0.95244561 0.95365554 0.95055187 0.95065711 0.95328733
 0.94260862 0.93419191 0.93103565 0.92808982 0.91351836 0.90568035
 0.91033411 0.91361421 0.92007852 0.92624009 0.93162614 0.93573272
 0.93819141 0.93902886]
8 day input [[0.93856746]]
9 day input [0.97080456 0.96785873 0.95854777 0.95838993 0.96885822 0.96743787
 0.96854261 0.95707483 0.96564937 0.97080456 0.97890565 0.99179373
 0.99105725 0.98884789 0.99384531 0.99147811 0.98826926 0.95896857
 0.94949978 0.9558649  0.94781646 0.94192473 0.94139869 0.95118316
 0.9389789  0.928195   0.93955754 0.94413416 0.93719034 0.95044668
 0.95244561 0.95365554 0.95055187 0.95065711 0.95328733 0.94260862
 0.93419191 0.93103565 0.92808982 0.91351836 0.90568035 0.91033411
 0.91361421 0.92007852 0.92624009 0.93162614 0.93573272 0.93819141
 0.93902886 0.93856746]
9 day input [[0.93728375]]
10 day input [0.96785873 0.95854777 0.95838993 0.96885822 0.96743787 0.96854261
 0.95707483 0.96564937 0.97080456 0.97890565 0.99179373 0.99105725
 0.98884789 0.99384531 0.99147811 0.98826926 0.95896857 0.94949978
 0.9558649  0.94781646 0.94192473 0.94139869 0.95118316 0.9389789
 0.928195   0.93955754 0.94413416 0.93719034 0.95044668 0.95244561
 0.95365554 0.95055187 0.95065711 0.95328733 0.94260862 0.93419191
 0.93103565 0.92808982 0.91351836 0.90568035 0.91033411 0.91361421
 0.92007852 0.92624009 0.93162614 0.93573272 0.93819141 0.93902886
 0.93856746 0.93728375]
10 day input [[0.935687]]
11 day input [0.95854777 0.95838993 0.96885822 0.96743787 0.96854261 0.95707483
 0.96564937 0.97080456 0.97890565 0.99179373 0.99105725 0.98884789
 0.99384531 0.99147811 0.98826926 0.95896857 0.94949978 0.9558649
 0.94781646 0.94192473 0.94139869 0.95118316 0.9389789  0.928195
 0.93955754 0.94413416 0.93719034 0.95044668 0.95244561 0.95365554
 0.95055187 0.95065711 0.95328733 0.94260862 0.93419191 0.93103565
 0.92808982 0.91351836 0.90568035 0.91033411 0.91361421 0.92007852
 0.92624009 0.93162614 0.93573272 0.93819141 0.93902886 0.93856746
 0.93728375 0.93568701]
11 day input [[0.9342226]]
12 day input [0.95838993 0.96885822 0.96743787 0.96854261 0.95707483 0.96564937
 0.97080456 0.97890565 0.99179373 0.99105725 0.98884789 0.99384531
 0.99147811 0.98826926 0.95896857 0.94949978 0.9558649  0.94781646
 0.94192473 0.94139869 0.95118316 0.9389789  0.928195   0.93955754
 0.94413416 0.93719034 0.95044668 0.95244561 0.95365554 0.95055187
 0.95065711 0.95328733 0.94260862 0.93419191 0.93103565 0.92808982
 0.91351836 0.90568035 0.91033411 0.91361421 0.92007852 0.92624009
 0.93162614 0.93573272 0.93819141 0.93902886 0.93856746 0.93728375
 0.93568701 0.93422258]
12 day input [[0.93320537]]
13 day input [0.96885822 0.96743787 0.96854261 0.95707483 0.96564937 0.97080456
 0.97890565 0.99179373 0.99105725 0.98884789 0.99384531 0.99147811
 0.98826926 0.95896857 0.94949978 0.9558649  0.94781646 0.94192473
 0.94139869 0.95118316 0.9389789  0.928195   0.93955754 0.94413416
 0.93719034 0.95044668 0.95244561 0.95365554 0.95055187 0.95065711
 0.95328733 0.94260862 0.93419191 0.93103565 0.92808982 0.91351836
 0.90568035 0.91033411 0.91361421 0.92007852 0.92624009 0.93162614
 0.93573272 0.93819141 0.93902886 0.93856746 0.93728375 0.93568701
 0.93422258 0.93320537]
13 day input [[0.93279195]]
14 day input [0.96743787 0.96854261 0.95707483 0.96564937 0.97080456 0.97890565
 0.99179373 0.99105725 0.98884789 0.99384531 0.99147811 0.98826926
 0.95896857 0.94949978 0.9558649  0.94781646 0.94192473 0.94139869
 0.95118316 0.9389789  0.928195   0.93955754 0.94413416 0.93719034
 0.95044668 0.95244561 0.95365554 0.95055187 0.95065711 0.95328733
 0.94260862 0.93419191 0.93103565 0.92808982 0.91351836 0.90568035
 0.91033411 0.91361421 0.92007852 0.92624009 0.93162614 0.93573272
 0.93819141 0.93902886 0.93856746 0.93728375 0.93568701 0.93422258
 0.93320537 0.93279195]
14 day input [[0.9329892]]
15 day input [0.96854261 0.95707483 0.96564937 0.97080456 0.97890565 0.99179373
 0.99105725 0.98884789 0.99384531 0.99147811 0.98826926 0.95896857
 0.94949978 0.9558649  0.94781646 0.94192473 0.94139869 0.95118316
 0.9389789  0.928195   0.93955754 0.94413416 0.93719034 0.95044668
 0.95244561 0.95365554 0.95055187 0.95065711 0.95328733 0.94260862
 0.93419191 0.93103565 0.92808982 0.91351836 0.90568035 0.91033411
 0.91361421 0.92007852 0.92624009 0.93162614 0.93573272 0.93819141
 0.93902886 0.93856746 0.93728375 0.93568701 0.93422258 0.93320537
 0.93279195 0.93298918]
15 day input [[0.93368775]]
16 day input [0.95707483 0.96564937 0.97080456 0.97890565 0.99179373 0.99105725
 0.98884789 0.99384531 0.99147811 0.98826926 0.95896857 0.94949978
 0.9558649  0.94781646 0.94192473 0.94139869 0.95118316 0.9389789
 0.928195   0.93955754 0.94413416 0.93719034 0.95044668 0.95244561
 0.95365554 0.95055187 0.95065711 0.95328733 0.94260862 0.93419191
 0.93103565 0.92808982 0.91351836 0.90568035 0.91033411 0.91361421
 0.92007852 0.92624009 0.93162614 0.93573272 0.93819141 0.93902886
 0.93856746 0.93728375 0.93568701 0.93422258 0.93320537 0.93279195
 0.93298918 0.93368775]
16 day input [[0.9347076]]
17 day input [0.96564937 0.97080456 0.97890565 0.99179373 0.99105725 0.98884789
 0.99384531 0.99147811 0.98826926 0.95896857 0.94949978 0.9558649
 0.94781646 0.94192473 0.94139869 0.95118316 0.9389789  0.928195
 0.93955754 0.94413416 0.93719034 0.95044668 0.95244561 0.95365554
 0.95055187 0.95065711 0.95328733 0.94260862 0.93419191 0.93103565
 0.92808982 0.91351836 0.90568035 0.91033411 0.91361421 0.92007852
 0.92624009 0.93162614 0.93573272 0.93819141 0.93902886 0.93856746
 0.93728375 0.93568701 0.93422258 0.93320537 0.93279195 0.93298918
 0.93368775 0.93470758]
17 day input [[0.9358484]]
18 day input [0.97080456 0.97890565 0.99179373 0.99105725 0.98884789 0.99384531
 0.99147811 0.98826926 0.95896857 0.94949978 0.9558649  0.94781646
 0.94192473 0.94139869 0.95118316 0.9389789  0.928195   0.93955754
 0.94413416 0.93719034 0.95044668 0.95244561 0.95365554 0.95055187
 0.95065711 0.95328733 0.94260862 0.93419191 0.93103565 0.92808982
 0.91351836 0.90568035 0.91033411 0.91361421 0.92007852 0.92624009
 0.93162614 0.93573272 0.93819141 0.93902886 0.93856746 0.93728375
 0.93568701 0.93422258 0.93320537 0.93279195 0.93298918 0.93368775
 0.93470758 0.93584841]
18 day input [[0.93692744]]
19 day input [0.97890565 0.99179373 0.99105725 0.98884789 0.99384531 0.99147811
 0.98826926 0.95896857 0.94949978 0.9558649  0.94781646 0.94192473
 0.94139869 0.95118316 0.9389789  0.928195   0.93955754 0.94413416
 0.93719034 0.95044668 0.95244561 0.95365554 0.95055187 0.95065711
 0.95328733 0.94260862 0.93419191 0.93103565 0.92808982 0.91351836
 0.90568035 0.91033411 0.91361421 0.92007852 0.92624009 0.93162614
 0.93573272 0.93819141 0.93902886 0.93856746 0.93728375 0.93568701
 0.93422258 0.93320537 0.93279195 0.93298918 0.93368775 0.93470758
 0.93584841 0.93692744]
19 day input [[0.9378108]]
[[0.9103341102600098], [0.9136142134666443], [0.9200785160064697], [0.926240086555481], [0.9316261410713196], [0.9357327222824097], [0.9381914138793945], [0.9390288591384888], [0.9385674595832825], [0.9372837543487549], [0.9356870055198669], [0.9342225790023804], [0.9332053661346436], [0.9327919483184814], [0.9329891800880432], [0.9336877465248108], [0.9347075819969177], [0.9358484148979187], [0.9369274377822876], [0.9378107786178589]]
In [76]:
len(df1)
Out[76]:
3943

Prediction¶

In [77]:
day_new=np.arange(1,51)
day_pred=np.arange(51,71)
In [78]:
df3=df1.tolist()
df3.extend(list_output)
In [79]:
plt.plot(day_new, Scaler.inverse_transform(df1[3893:]))
plt.plot(day_pred, Scaler.inverse_transform(list_output))
Out[79]:
[<matplotlib.lines.Line2D at 0x2bf1f54db80>]
In [80]:
df3=df1.tolist()
df3.extend(list_output)
plt.plot(df3[3800:])
Out[80]:
[<matplotlib.lines.Line2D at 0x2bf1f4fd820>]
In [81]:
df3=Scaler.inverse_transform(df3).tolist()
In [82]:
plt.plot(df3)
Out[82]:
[<matplotlib.lines.Line2D at 0x2bf1f5bf520>]
In [83]:
x_input=test_data[len(test_data)-time_step:].reshape(1,-1)
temp_input=list(x_input)
temp_input=temp_input[0].tolist()

from numpy import array

lst_output=[]
n_steps=time_step
i=0
pred_days = 20
while(i<pred_days):
    
    if(len(temp_input)>time_step):
        
        x_input=np.array(temp_input[1:])
        #print("{} day input {}".format(i,x_input))
        x_input = x_input.reshape(1,-1)
        x_input = x_input.reshape((1, n_steps, 1))
        
        yhat = model.predict(x_input, verbose=0)
        #print("{} day output {}".format(i,yhat))
        temp_input.extend(yhat[0].tolist())
        temp_input=temp_input[1:]
        #print(temp_input)
       
        lst_output.extend(yhat.tolist())
        i=i+1
        
    else:
        
        x_input = x_input.reshape((1, n_steps,1))
        yhat = model.predict(x_input, verbose=0)
        temp_input.extend(yhat[0].tolist())
        
        lst_output.extend(yhat.tolist())
        i=i+1
               
print("Output of predicted next days: ", len(lst_output))

Output of predicted next days:  20
In [84]:
last_days=np.arange(1,time_step+1)
day_pred=np.arange(time_step+1,time_step+pred_days+1)
print(last_days)
print(day_pred)

[ 1  2  3  4  5  6  7  8  9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48
 49 50]
[51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70]
In [85]:
import plotly.express as px
from itertools import cycle
In [86]:
temp_mat = np.empty((len(last_days)+pred_days+1,1))
temp_mat[:] = np.nan
temp_mat = temp_mat.reshape(1,-1).tolist()[0]

last_original_days_value = temp_mat
next_predicted_days_value = temp_mat

last_original_days_value[0:time_step+1] = Scaler.inverse_transform(df1[len(df1)-time_step:]).reshape(1,-1).tolist()[0]
next_predicted_days_value[time_step+1:] = Scaler.inverse_transform(np.array(lst_output).reshape(-1,1)).reshape(1,-1).tolist()[0]

new_pred_plot = pd.DataFrame({
    'last_original_days_value':last_original_days_value,
    'next_predicted_days_value':next_predicted_days_value
})
names = cycle(['Last 50 days close price','Predicted next 30 days close price'])

fig = px.line(new_pred_plot,x=new_pred_plot.index, y=[new_pred_plot['last_original_days_value'],
                                                      new_pred_plot['next_predicted_days_value']],
              labels={'value': 'Stock price','index': 'Timestamp'})
fig.update_layout(title_text='Compare last 50 days vs next 20 days',
                  plot_bgcolor='white', font_size=15, font_color='black', legend_title_text='Close Price')
fig.for_each_trace(lambda t:  t.update(name = next(names)))

fig.update_xaxes(showgrid=False)
fig.update_yaxes(showgrid=False)
fig.show()
In [87]:
lstmgrudf=df1.tolist()
lstmgrudf.extend((np.array(lst_output).reshape(-1,1)).tolist())
lstmgrudf=Scaler.inverse_transform(lstmgrudf).reshape(1,-1).tolist()[0]

names = cycle(['Close price'])

fig = px.line(lstmgrudf,labels={'value': 'Stock price','index': 'Timestamp'})
fig.update_layout(title_text='Plotting whole closing stock price with prediction',
                  plot_bgcolor='white', font_size=15, font_color='black', legend_title_text='Stock')
fig.for_each_trace(lambda t:  t.update(name = next(names)))

fig.update_xaxes(showgrid=False)
fig.update_yaxes(showgrid=False)
fig.show()

Screenshot%202023-11-20%20162819.png

Insights¶

-